The impact of age mimicry bias on the accuracy of methods for age estimation based on Kvaal's pulp/tooth ratios: a bootstrap study.

Introduction: Methods based on pulp/tooth ratios proposed by Kvaal et al. (1995) have been widely used for age estimation in adults. The tendency of age estimates to mimic the age structure of the reference population, i.e. age mimicry bias, is a possible source of controversy in the results of studies. The primary objective of the present study was to evaluate the effect of age mimicry bias on the accuracy of the original Kvaal's method and its recent modification proposed by Roh et al. (2018).

Method And Material: The study sample comprised 240 cone-beam computed tomography scans of an Iranian population. The bootstrap procedure was used to study the impact of age mimicry bias on age estimates by constructing reference populations with different age structures. The accuracy of Kvaal's and Roh's variables for age estimation was assessed using a twofold cross-validation technique and principal component analysis.

Results: The application of original equations resulted in SEE values highly greater than the acceptable threshold for forensic purposes (10 years). The population-specific equations obtained SEE values of 9.18, 11.03, and 9.22 years using Kvaal's variables and 9.19, 11.13, and 9.14 years using Roh's variable for the maxillary, mandibular, and all teeth, respectively. The bootstrap procedure revealed that using uneven reference populations to formulate the equation resulted in significantly greater SEEs (almost all >10 years) that were not acceptable for forensic purposes. Moreover, using an older (a younger) reference population contributed to a sharp rise in the amount of over- (under-) estimation for younger (older) age groups.

Conclusion: Age mimicry bias had an undesirable impact on the accuracy of dental age estimation based on Kvaal's and Roh's methods. The accuracy of the original equations was low for age estimation in the Iranian sample. However, the performance of the population-specific equations was reasonably acceptable for the maxillary and all six teeth.